The goal of this project is to use fMRI brain imaging to understand the organization and acquisition of complex cognitive abilities. Components of the ACT-R cognitive architecture, which is capable of modeling complex cognitive abilities, have been mapped on different brain regions. A methodology has been developed for taking the activities of these components and making predictions for the BOLD response obtained in these areas from fMRI imaging. Recent developments have also enabled such tests to be applied in the face of high variability in the timing of individual components. The proposed research is intended to further develop the mapping between components of the ACT-R theory and to extend the work to understand the structure of complex tasks. Three classes of experiments are proposed. First, experiments will be performed to test predictions about a prefrontal region associated with declarative retrieval, a parietal region associated with mental representation, and an anterior cingulate region associated with control. The second group of experiments are designed to extend and modify the ACT-R theory. They will examine whether the response of an anterior prefrontal region can guide the development of a metacognitive component in ACT-R and whether the ACT-T theory of the dorsal caudate should be amended to include effects of reinforcement learning. The third class of experiments will test the decomposition hypothesis -- that the processes in large task consist of the same components as are revealed in smaller tasks. A complex radar-screen task will be investigated to see if the activation patterns in this task can be predicted from the behavior of individual components of the ACT-R theory. It will also be studied to identify further directions for the development of the cognitive architecture. In addition to advancing understanding of complex cognition this research will serve to develop the methodology for using fMRI research to study higher-level cognitive functioning. Advancing the mapping of fMRI brain imaging to higher-level cognitive function is important for many applications including health-related efforts such as understanding the basis of cognitive dysfunctions. For instance, many theories of autism relate it to the coordination of different brain regions in service of intellectual goals.